DSXchange

Hi,
Pls let me know, how to update the source table in the same job based on some condition.

Why bother with a DataStage job? All you need is an UPDATE statement in the database.

As noted, sounds like a job for SQL. If you disagree, please expand on your question. Alot.

If you do this then you might lock the records you are trying to update. So the source SQL needs to have the correct options on it.

kduke wrote:If you do this then you might lock the records you are trying to update. So the source SQL needs to have the correct options on it.

This is my concern. We are in desing phase, and dont have access to Datastage. So kindly let me know, if I read some data from a table, and trying to updated the same records with some change, will there be any lock. Can we acheive with read commit and Commit interval of 1 record?

I think you want read uncommitted but I need to look it up. Sometimes you can do it in user defined SQL with a NOLOCK or whatever the database needs based on syntax.

Unload your table into a flat file or dataset and then do your transformations and loading.

DSguru2B wrote:Unload your table into a flat file or dataset and then do your transformations and loading.

But it again require two jobs, or some superfluous flow isolating stages.

My direction was towards modularization. More like an advice on not to do it in a single job.

Yes I agree with DSGuru2B. In my project we have this kind of scenario. We always use 2 jobs and it works fine.
1. Dump the tables into dataset for reading (job 1)
2. Update the same table by lookup on the dataset (job 2)

So, if we can have 79, is 2B a hex number. And, if so, is 79 a hex number?

Very funny Ray, 2B a bionic number.

And here's me assuming it was a blackness grade in pencils...

I think doing this in one job is fine.

Thanks Kim.

May I know how to specify the Uncommited read?